This node looks like it needs updating on the current state of play with quantum computing. For a start, of course, we still don't have one. There are no useful quantum computers. Quantum software is coming along nicely, so if we create one, it won't just sit there unused for years.
You'd know if we had a full-scale one. Oh yes. As mentioned above, the RSA encyption algorithm is dependent upon factorization being very hard to do. Quantum computers make it quick and easy. There is quantum encyption, of course, but that's a seperate problem and is not likely to be implemented any sooner than quantum computing renders our current codes redundant.
So, how is quantum computing actually coming along?
Physicists hijacked a medical method called nuclear magnetic resonance, which allows nuclei to be given a certain spin representing a qubit using electromagnetic fields - typically from radio waves. Nuclei cunningly dodge the decoherence problem described above because they don't interact with their environment much.
This seemed promising for a while. IBM built a 7-qubit device, and successfully factored 15 with it. There have been talks of a 10-qubit computer, but sadly that's nearing the upper limit for NMR. Scientists don't believe that a machine of more than fifteen qubits is possible with this method.
The reasons for this are two-fold. For a start, the computer's return is recorded by measuring the magnetic field of the nuclei. These are so weak that you need a lot of molecules with the same spin to make it measurable. This could be theoretically overcome, but the real problem is that, unlike traditional computing, you don't know the initial spin of the molecules involved. This noise completely overcomes the signal once you have enough molecules to represent more than fifteen or so qubits.
Recent efforts have turned to that great bastion of engineering, solid state technology. Promoted by David DeVincenzo, this method uses so-called "quantum dots" on a silicon chip, controlled by electrodes. So far, only one qubit devices have been constructed, but importantly, there's no obvious signs of an upper limit so far. We could have several qubits on one chip in a few years. One to watch.
A second possibility is a superconducting circuit, which uses superposition to make the current flow both ways. These could be small and easily placed onto chips, creating large-scale circuits like those of a traditional computer. So far, there's been no success in linking these chips together, but again, there's no suggestion as yet that this is a dead-end.
So far then, we have a few avenues for research, and numerous other contenders by the way side. There's no theoretical reason why a useful quantum computer can't be created, as far as we know. This puts physicists in a "win-win" situation; if a quantum computer is built, all well and good. If they can find some reason why it cannot, as a matter of principle, ever be built, they have a whole new area of theory to explore. Most physicists are nevertheless hopeful that it can be achieved. With any luck, it's just a matter of time.